摘要
针对传统烧结配料方法准确率和经济性不高的问题,从烧结技术要求和经济效益两方面综合考虑,提出基于预测模型与调整规则的烧结配料优化综合集成方法。根据烧结试验和生产所获得的工业数据建立成分预测数学模型,以群体专家经验得到的定性知识构造配比调整规则模型,利用线性规划和神经网络方法建立烧结配料优化模型和烧结矿性能指标预测模型,采用从定性到定量综合集成方法,将这些模型综合集成,从而实现烧结配料的优化,应用结果验证了该方法的有效性。
To deal with the problem of low accuracy and high cost existed in traditional methods of sinter mix proportions, an integrated synthesis methodology was proposed based on predictive models and adjustment rules. First, mathematical proportion-prediction models were derived on the basis of sinter experiments and data of actual runs. And adjustment rules were extracted from qualitative knowledge of experts＇ experiences. Then, a sinter mix proportion model for optimization and a predictive model for sinter quality were built by integrating the derived models using the methods of linear programming and neural network. Finally, the proportions were optimized through an integrated synthesis methodology under both qualitative and quantitative considerations. Application results demonstrate the validity of this methodology.
出处
《系统仿真学报》
EI
CAS
CSCD
北大核心
2008年第9期2423-2428,共6页
Journal of System Simulation
基金
国家杰出青年科学基金项目(60425310)
关键词
烧结过程
配料优化
预测模型
调整规则
定性定量综合集成
sintering process
sinter mix proportion optimization
prediction model
adjustment rule
meta-synthesis
作者简介
吴敏（1963-），男，广东化州人，长江学者特聘教授，博导，研究方向为过程控制，鲁棒控制和智能系统。
王春生（1966-），男，河南禹州人，博士生，研究方向为复杂系统建模与优化控制。
